Recovering from offline volumes

You can use the management GUI or issue commands to recover from an offline volume after a node or an I/O group has failed.

About this task

If you lose both nodes in an I/O group, you lose access to data and all volumes that are associated with the I/O group. To regain access to the volumes, you must complete one of the following procedures. Depending on the failure scenario, cached data for these volumes might be lost and the volumes are now offline.

Procedure

Data loss scenario 1

One node in an I/O group failed and the failover process has started on the second node. During the failover process, the second node in the I/O group fails before the data in the write cache is written to the backend. The first node is successfully repaired. However, because the hardened data is not the most recent version that is committed to the data store, it cannot be used. The second node is replaced but has lost data. Therefore, the node cannot recognize that it is part of the cluster.

Complete steps 1 through 4 to recover from offline volumes when one node has earlier-level hardened data and the other node has lost hardened data.

  1. Recover the node and add it back into the cluster.
  2. Delete all FlashCopy®, Metro Mirror, and Global Mirror mappings and relationships that use the offline volumes.
  3. Run the recovervdisk, recovervdiskbyiogrp, or recovervdiskbycluster command.
  4. Re-create all FlashCopy, Metro Mirror, and Global Mirror mappings and relationships that use the volumes.

Data loss scenario 2

Both nodes in the I/O group failed and were repaired. Because the nodes have lost their hardened data, they have no way of recognizing that they are part of the cluster.

Complete steps 5 through 7 to recover from offline volumes when both nodes lose their hardened data and cannot be recognized by the cluster.

  1. Delete all FlashCopy, Metro Mirror, and Global Mirror mappings and relationships that use the offline volumes.
  2. Run the recovervdisk, recovervdiskbyiogrp, or recovervdiskbycluster command.
  3. Re-create all FlashCopy, Metro Mirror, and Global Mirror mappings and relationships that use the volumes.
Parent topic: Servicing